Meta-analysis
Pooling of effect estimates across studies, with assessment of heterogeneity, sensitivity and publication bias.
Random-effects models, forest and funnel plots, subgroup analysis
Rigorous evidence synthesis, statistical analysis and systematic-review support for researchers across medicine, life sciences, management, economics, psychology, education and the social sciences.
My Meta-Analysis is a research-support organization specializing in meta-analysis, systematic reviews and statistical analysis.
We support the full evidence-synthesis workflow: protocol, search, screening, extraction, analysis, reporting and journal submission.
Researchers and research teams preparing reviews and analyses for publication, from clinical medicine to management and economics.
Each service follows the relevant published methodological standards and is scoped to the study design, not to a template.
Pooling of effect estimates across studies, with assessment of heterogeneity, sensitivity and publication bias.
Random-effects models, forest and funnel plots, subgroup analysis
A documented search, screening, appraisal and synthesis, reported against PRISMA 2020.
Protocol, search strategy, risk of bias, flow diagram
Comparison of several interventions at once, using direct and indirect evidence.
Transitivity, consistency, treatment rankings, league tables
Synthesis with explicit prior assumptions, useful when few studies are available or prior evidence should be formally incorporated.
Prior specification, posterior intervals, sensitivity to priors
Examination of whether study-level characteristics explain differences between results.
Moderators, mixed-effects models, limits of study-level inference
Design-appropriate analysis of primary research data, with reproducible code and a clear methods description.
Model selection, survival analysis, sample-size considerations
Reporting-guideline checks, manuscript editing, journal selection and responses to reviewer comments.
PRISMA and MOOSE adherence, submission materials, revisions
From pairwise meta-analysis to network and Bayesian models, and from systematic reviews to scoping and qualitative synthesis.
Not every method suits every field or dataset. Choosing the right design, or recommending against pooling, is part of the work.
Medicine and health sciences are a flagship area. The same methods apply, with discipline-specific study designs and outcomes, across the fields below.
Clinical research, public health, epidemiology, nursing, pharmacy, oncology, cardiology, neurology, psychiatry
Genetics, genomics, microbiology, immunology, neuroscience, ecology, evolutionary biology
Clinical, cognitive, social, developmental and educational psychology; mental health and addiction research
Organizational behavior, human resource management, marketing, entrepreneurship, supply chain
Health, development, labor, environmental and behavioral economics; economic policy
Higher education, STEM education, medical education, online learning, special education
Sociology, social work, political science, public policy, criminology, migration studies
Climate change, conservation, biodiversity, water resources, environmental health
Agronomy, crop and soil science, animal and veterinary research, food science
Exercise physiology, sports medicine, physical activity, injury prevention, rehabilitation
Biomedical, environmental and energy engineering; computer science and machine learning
A synthesis is only as useful as the method behind it. These are the commitments the work is organized around.
The review design and statistical model follow from the research question and the available studies, not from a fixed package.
Reviews are planned and reported against recognized guidance, including the Cochrane Handbook, PRISMA 2020 and the relevant extensions.
We provide research and evidence-synthesis support. We do not provide clinical advice, and results are not patient-specific guidance.
Our research integrity and authorship statement and AI-use policy set out how work is done and credited.
Most projects move through five stages. Some, such as a statistical re-analysis for a revision, need only one or two.
We review your question, study type, target journal and deadline, and agree what is feasible.
Eligibility criteria, outcomes and analysis plan are specified, and registration is prepared where appropriate.
Searches are designed and documented, studies screened, and data extracted using a defined form.
Risk of bias and certainty of evidence are assessed, and the pre-specified analyses are run.
Results are written up against the relevant reporting guideline, with support through submission and revision.
What you receive depends on the service. A full review project can include:
Evidence synthesis should be checkable by someone else. The working practices below are designed to make that possible.
Short, referenced explanations of the concepts that come up most often when planning and reporting an evidence synthesis.
A statistical method for combining quantitative results from separate studies that address the same question, to estimate an overall effect and examine why results differ.
A fixed-effect model assumes one true effect shared by every study. A random-effects model assumes true effects vary, and estimates their average and spread.
A reporting guideline for systematic reviews, consisting of a 27-item checklist and a flow diagram that documents how studies were identified and selected.
Answers to common questions about reviews and meta-analyses. More are in the full FAQ.
A systematic review is a structured method for identifying, appraising and summarizing the relevant studies on a question. A meta-analysis is a statistical technique for combining the numerical results of those studies. Many systematic reviews include a meta-analysis, but a review can be reported narratively when the studies are too different to pool.
In most cases it should be. A meta-analysis is normally built on a systematic, documented search and selection of studies. Pooling studies chosen informally risks biased estimates, because the result then depends on which studies happened to be found.
There is no universal minimum. Two studies can be combined mathematically, but with few studies the estimate of between-study variation is imprecise, and methods such as meta-regression or tests for publication bias become unreliable. Whether pooling is meaningful depends on how similar the studies' questions are as much as on how many there are.
PRISMA 2020 is the standard reporting guideline for systematic reviews, with extensions for scoping reviews, diagnostic accuracy reviews, network meta-analyses and other designs. Some journals also ask for MOOSE when a meta-analysis is based on observational studies. The target journal's instructions for authors decide which applies.
No. It requires studies that measure comparable outcomes in a comparable way. It is well established in clinical medicine, psychology and education, and used in parts of the social, management and ecological sciences, but it is less common elsewhere. In some fields a systematic or scoping review without pooling is the more appropriate design.
No. My Meta-Analysis provides research and evidence-synthesis support. The results of a review or analysis are not patient-specific medical advice.
Describe your question, study type and target journal. We will respond with the approach we would recommend and what we would need to begin.